Applying this classification to occupational employment data, we \nestimate that 37% of dependent employment in the EU27 can technically be\n carried out remotely. This estimate is very close to those indicated in\n real-time surveys conducted during the Covid-19 crisis, notably \nEurofound\u2019s \u201cLiving, Working and Covid-19\u201d e-survey (Eurofound 2020). \nOur estimates of the fraction of teleworkable employment range from 35% \nto 41% in two thirds of EU countries, with the highest value in \nLuxembourg (54%) and the lowest in Romania (27%) (see Figure 1).<\/p>\n\n\n\n
Figure 1<\/strong> Percentage of employees in teleworkable occupations by member state, EU27<\/p>\n\n\n\n Note<\/em>: Employees only. Overall, these estimates likely provide an \u2018upper bound\u2019 on the \npercentage of jobs that can be done remotely in an efficient way. The \nmajority of teleworkable jobs requires extensive social interaction, \nwhich often make working remotely sub-optimal. Even the most \nsophisticated videoconference systems are unlikely to match the quality \nof face-to-face interactions (e.g. Schoenenberg et al. 2014), be it for \nmedical advice, counselling, teaching, and so on.<\/p>\n\n\n\n On this basis, we estimate that only 13% of employment in Europe is \nin teleworkable occupations that involve no or limited social tasks \n(e.g. selling, teaching, caring for others, working with the public) and\n can in principle be carried out remotely with no or limited loss of \nquality. The remaining 24% of technically teleworkable jobs involve \nextensive social interaction and thus they can only be partially \nprovided remotely (for instance, for some but not all tasks) without a \nsignificant loss in service quality.<\/p>\n\n\n\n When breaking down figures of technical teleworkability by broad \noccupation group (see Figure 2), a first striking difference emerges \nbetween white-collar and blue-collar work, with the latter being much \nless teleworkable mainly due to the physical requirements of the jobs \nand the associated place dependence.<\/p>\n\n\n\n Another very interesting result concerns the occupational \ndistribution of the gap between actual and potential teleworking, with \nthe highest figure observed for lower-level white-collar occupations. \nIndeed, although the characteristics of the job are such that almost all\n (84%) clerical support workers could telework, only 5% of them had \nworked from home before the Covid-19 crisis. Such findings would suggest\n that, beyond technical feasibility, differences in access to \nteleworking across occupations also depend on aspects related to the \norganisation of work and the position in the occupational hierarchy (and\n related privileges), rather than the task composition of the job as \nsuch. As already argued in an earlier Vox column (Fern\u00e1ndez-Mac\u00edas and \nBisello 2016), work organisation is a crucial aspect of the task profile\n of occupations which can also affect access to teleworking in \nparticular cases. The fact that managers typically enjoy more autonomy \nand are subject to less monitoring of their work effort than secretaries\n is probably related to the much larger prevalence of teleworking for \nmanagers before the Covid-19 crisis, even if the work of secretaries is \ntechnically more teleworkable. The sudden extension of teleworking to \noccupations where it was previously less prevalent is probably inducing \nimportant changes in work organisation, including a growing utilisation \nof digital tools for the remote control and monitoring of work effort.<\/p>\n\n\n\n Figure 2<\/strong> Teleworkability and actual teleworking as a share of employment by broad occupation group, EU27<\/p>\n\n\n\n Note<\/em>: Employees only. When looking at the socioeconomic profile of workers in teleworkable \noccupations, stark differences emerge between high- and low-paid \nworkers, with three-quarters (74%) of those in jobs in the highest \npaying wage quintile who can telework, compared to only 3% of those in \nthe lowest quintile (see Figure 3). A divide is also seen when looking \nat the differentiation by educational qualification, with around 66% of \ntertiary education graduates working in teleworkable occupations, \nagainst a much smaller share of those with lower levels of \nqualification. Gender differences also emerge, with a much higher share \nof women than men (45% compared to 30%) in teleworkable occupations. \nThis reflects to some extent to patterns of sectoral segregation \u2013 as \nwomen are under-represented in sectors such as agriculture, mining, \nmanufacturing, utilities and construction with limited teleworkability \u2013\n but also differences in the job profiles even within male-dominated \nsectors, with women being more likely to be in office-based, secretarial\n or administrative jobs which are more amenable to remote working.<\/p>\n\n\n\n Meanwhile, more than 40% of employees living in cities are in \nteleworkable occupations, against fewer than 30% of those living in \nrural areas, which reflects the fact that cities have larger fraction of\n employment in knowledge- and ICT-intensive occupations than towns or \nrural areas. Employees in medium and large firms are also significantly \nmore likely to be in teleworkable occupations than those working in \nmicro-enterprises. <\/p>\n\n\n\n Figure 3<\/strong> Employees in teleworkable occupations by workers\u2019 characteristics, EU-27 (%)<\/p>\n\n\n\n Note<\/em>: Employees only. Job-wage quintiles based on author\u2019s calculations of SES 2014 data. What are the implications of the ongoing ad hoc experiment in mass \nteleworking for the future of work? Evidence from our study suggests \nthat access to teleworking during the pandemic has become more evenly \ndistributed within white-collar occupations, resulting in new \npossibilities amongst low- and mid-level clerical and administrative \nworkers. In teleworking, the monitoring of work effort is much more \ndifficult and thus requires a higher level of trust. In this sense, the \nexpansion of teleworking could shift cultural and organisational norms, \nexpanding levels of work autonomy and making more accessible what so far\n has often been a privilege associated with high professional status. \nYet, firms\u2019 and workers\u2019 lack of affinities with digital tools and prior\n experience with remote working arrangements may limit its uptake and \neffectiveness (Milasi et al. 2020). There is also the danger that \norganisations respond to this challenge by using intrusive digital tools\n for remotely monitoring work effort, which has troubling implications \nin terms of job quality, privacy and autonomy.<\/p>\n\n\n\n At the same time, a new divide between those who can telework and \nthose who cannot has emerged. Indeed, dramatic differences can be \nobserved in teleworkability by wages and by education level. Experienced\n employees in white-collar occupations, often in knowledge-based \nservices, reported the highest incidence of teleworking during Covid-19.\n In order to avoid a teleworking divide, access to remote working \narrangements should be facilitated also among younger and \nlower-qualified employees. As workers with strong digital skills are \narguably better positioned to respond to the demands of remote working \nduring the current crisis, it is important that extensive training \nopportunities are offered. However, as long as the bottleneck of \nphysical tasks remains (jobs with significant manual task content cannot\n be teleworked), a broad expansion of telework would unavoidably \nincrease the social distance between those that need to work with their \nhands in particular places and those that can provide intellectual and \nsocial services from any place.<\/p>\n\n\n\n Berg, J, F Bonnet and S Soares (2020), \u201cWorking from home: Estimating the worldwide potential<\/a>\u201d, VoxEU.org, 11 May.<\/p>\n\n\n\n Bisello, M and E Fernandez-Macias (2020), \u201cA Taxonomy of Tasks for Assessing the Impact of New Technologies on Work<\/a>\u201d, JRC Working Papers Series on Labour, Education and Technology 2020\/04, JRC120618, European Commission.<\/p>\n\n\n\n Dingel, J and B Neiman (2020), \u201cHow many jobs can be done at home?<\/a>\u201d, VoxEU.org, 7 April.<\/p>\n\n\n\n Eurofound (2020), Living, working and COVID-19: First findings \u2013 April 2020<\/em><\/a>, Dublin.<\/p>\n\n\n\n European Commission (2020), 2020 European Semester: Country-specific recommendations<\/em>.<\/a><\/p>\n\n\n\n Fernandez-Macias, E and M Bisello (2016), \u201cA framework for measuring tasks across occupations<\/a>\u201d, VoxEU.org, 25 September.<\/p>\n\n\n\n Milasi S, I Gonz\u00e1lez-V\u00e1zquez and E Fernandez-Macias (2020), \u201cTelework in the EU before and after the COVID-19: where we were, where we head to<\/a>\u201d, JRC Science for Policy Brief<\/em>.<\/p>\n\n\n\n Schoenenberg, K, A Raake and J Koeppe (2014), “Why are you so slow? \nMisattribution of transmission delay to attributes of the conversation \npartner at the far-end\u201d, International Journal of Human-Computer Studies <\/em>72(5): 477-487.<\/p>\n\n\n\n![\"\"\/](\"https:\/\/voxeu.org\/sites\/default\/files\/image\/FromMay2014\/MilasiFigure1.PNG\")
<\/figure>\n\n\n\n
Source<\/em>: Authors\u2019 calculations from EU LFS.<\/p>\n\n\n\nWhat jobs can be done from home?<\/h2>\n\n\n\n
![\"\"\/](\"https:\/\/voxeu.org\/sites\/default\/files\/image\/FromMay2014\/MilasiFigure2.PNG\")
<\/figure>\n\n\n\n
Source<\/em>: Authors\u2019 calculations from EU LFS.<\/p>\n\n\n\n![\"\"\/](\"https:\/\/voxeu.org\/sites\/default\/files\/image\/FromMay2014\/MilasiFigure3.PNG\")
<\/figure>\n\n\n\n
Source<\/em>: Authors\u2019 calculations based on EU LFS and Structure of Earnings Survey.<\/p>\n\n\n\nPost-Covid-19 teleworking patterns: A new digital divide?<\/h2>\n\n\n\n
References<\/h2>\n\n\n\n